The effects of bariatric surgery in patients with nonalcoholic fatty liver disease (NASH) are not well established. We performed a prospective study to determine the biological and clinical effects of bariatric surgery in patients with NASH.

From May 1994 through May 2013, one hundred and nine morbidly obese patients with biopsy-proven NASH underwent bariatric surgery at the University Hospital of Lille, France (the Lille Bariatric Cohort). Clinical, biological, and histologic data were collected before and 1 year after surgery.

One year after surgery, NASH had disappeared from 85% of the patients (95% confidence interval [CI]: 75.8%−92.2%). Compared with before surgery, patients had significant reductions in mean ± SD body mass index (BMI, from 49.3 ± 8.2 to 37.4 ± 7) and level of alanine aminotransferase (from 52.1 ± 25.7 IU/L to 25.1 ± 20 IU/L); mean levels of γ-glutamyltransferases were reduced from 51 IU/L before surgery (interquartile range [IQR], 34−87 IU/L) to 23 IU/L afterward (IQR, 14−33 IU/L) and mean insulin resistance index values were reduced from 3.6 ± 0.5 to 2.9 ± 0.5 (P < .01 for each comparison). NASH disappeared from a higher proportion of patients with mild NASH before surgery (94%) than severe NASH (70%) (P < .05) according to Brunt score. In histologic analysis, steatosis was detected in 60% of the tissue before surgery (IQR, 40%−80%) but only 10% 1 year after surgery (IQR, 2.5%−21.3%); the mean nonalcoholic fatty liver disease score was reduced from 5 (IQR, 4−5) to 1 (IQR, 1−2) (each P < .001). Hepatocellular ballooning was reduced in 84.2% of samples (n = 69; 95% CI: 74.4−91.3) and lobular inflammation in 67.1% (n = 55; 95% CI: 55.8−77.1). According to Metavir scores, fibrosis was reduced in 33.8% of patients (95% CI: 23.6%−45.2%). Patients whose NASH persisted 1 year after surgery (n = 12) had lost significantly less weight (change in BMI, 9.1 ± 1.5) than those without NASH (change in BMI, 12.3 ± 0.6) (P = .005). Patients who underwent laparoscopic gastric banding lost less weight (change in BMI, 6.4 ± 0.7) than those who underwent gastric bypass (change in BMI, 14.0 ± 0.5) (P < .0001), and a higher proportion had persistent NASH (30.4% vs 7.6% of those with gastric bypass; P = .015).

Bariatric surgery induced the disappearance of NASH from nearly 85% of patients and reduced the pathologic features of the disease after 1 year of follow-up. It could be a therapeutic option for appropriate morbidly obese patients with NASH who do not respond to lifestyle modifications. More studies are needed to determine the long-term effects of bariatric surgery in morbidly obese patients with NASH.